From: Mark Lord <lkml@rtr.ca>
To: Jens Axboe <axboe@suse.de>,
IDE/ATA development list <linux-ide@vger.kernel.org>,
Linux Kernel <linux-kernel@vger.kernel.org>,
linux-scsi <linux-scsi@vger.kernel.org>
Subject: QUEUE_FLAG_CLUSTER: not working in 2.6.24 ?
Date: Thu, 13 Dec 2007 13:36:02 -0500 [thread overview]
Message-ID: <47617B92.6020908@rtr.ca> (raw)
Jens,
I'm experimenting here with trying to generate large I/O through libata,
and not having much luck.
The limit seems to be the number of hardware PRD (SG) entries permitted
by the driver (libata:ata_piix), which is 128 by default.
The problem is, the block layer *never* sends an SG entry larger than 8192 bytes,
and even that size is exceptionally rare. Nearly all I/O segments are 4096 bytes,
so I never see a single I/O larger than 512KB (128 * 4096).
If I patch various parts of block and SCSI, this limit doesn't budge,
but when I change the hardware PRD limit in libata, it scales by exactly
whatever I set the new value to. This tells me that adjacent I/O segments
are not being combined.
I thought that QUEUE_FLAG_CLUSTER (aka. SCSI host .use_clustering=1) should
result in adjacent single pages being combined into larger physical segments?
This is x86-32 with latest 2.6.24-rc*.
I'll re-test on older kernels next.
???
next reply other threads:[~2007-12-13 18:36 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-12-13 18:36 Mark Lord [this message]
2007-12-13 18:37 ` QUEUE_FLAG_CLUSTER: not working in 2.6.24 ? Mark Lord
2007-12-13 18:42 ` Matthew Wilcox
2007-12-13 18:46 ` James Bottomley
2007-12-13 18:48 ` Mark Lord
2007-12-13 18:53 ` Matthew Wilcox
2007-12-13 19:03 ` Mark Lord
2007-12-13 19:26 ` Jens Axboe
2007-12-13 19:30 ` Mark Lord
2007-12-13 19:32 ` Mark Lord
2007-12-13 19:39 ` Jens Axboe
2007-12-13 19:42 ` Mark Lord
2007-12-13 19:53 ` Jens Axboe
2007-12-13 19:59 ` Mark Lord
2007-12-13 20:05 ` Jens Axboe
2007-12-13 20:02 ` Jens Axboe
2007-12-13 20:06 ` Mark Lord
2007-12-13 20:09 ` Jens Axboe
2007-12-13 20:14 ` Mark Lord
2007-12-13 20:18 ` Mark Lord
2007-12-13 20:21 ` Jens Axboe
2007-12-13 22:02 ` Andrew Morton
2007-12-13 22:15 ` James Bottomley
2007-12-13 22:29 ` Andrew Morton
2007-12-13 22:33 ` Mark Lord
2007-12-13 23:13 ` Mark Lord
2007-12-14 0:05 ` Mark Lord
2007-12-14 0:30 ` Mark Lord
2007-12-14 0:37 ` Andrew Morton
2007-12-14 0:42 ` Mark Lord
2007-12-14 0:46 ` [PATCH] fix page_alloc for larger I/O segments (improved) Mark Lord
2007-12-14 0:57 ` James Bottomley
2007-12-14 1:11 ` Andrew Morton
2007-12-14 2:23 ` Mark Lord
2007-12-14 17:42 ` Mel Gorman
2007-12-14 18:07 ` Mark Lord
2007-12-16 21:56 ` Mel Gorman
2007-12-14 18:13 ` Matthew Wilcox
2007-12-14 18:30 ` Mark Lord
2007-12-20 22:37 ` Matthew Wilcox
2007-12-14 0:47 ` QUEUE_FLAG_CLUSTER: not working in 2.6.24 ? Mark Lord
2007-12-14 11:50 ` Mel Gorman
2007-12-14 13:57 ` Mark Lord
2007-12-14 0:40 ` [PATCH] fix page_alloc for larger I/O segments Mark Lord
2007-12-14 1:03 ` Andrew Morton
2007-12-14 4:00 ` Matthew Wilcox
2007-12-13 22:17 ` QUEUE_FLAG_CLUSTER: not working in 2.6.24 ? Jens Axboe
2007-12-13 22:02 ` VM allocates pages in reverse order again Matthew Wilcox
2007-12-13 19:37 ` QUEUE_FLAG_CLUSTER: not working in 2.6.24 ? Jens Axboe
2007-12-13 19:53 ` Mark Lord
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=47617B92.6020908@rtr.ca \
--to=lkml@rtr.ca \
--cc=axboe@suse.de \
--cc=linux-ide@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).